The effect of fungal density on fungicide dose–response curves in barley powdery mildew ( Erysiphe graminis f.sp. hordei )

The effect of inoculation density on fungicide dose–response curves and the estimated ED 50 value (the fungicide concentration needed to halve the infected leaf area relative to the infected area of leaves that had not been treated with a fungicide) were investigated theoretically, and predictions were tested experimentally using powdery mildew of barley ( Erysiphe graminis [syn. Blumeria graminis ] f.sp. hordei ). A host leaf was assumed to consist of a number of independent compartments, in each of which only a single fungal spore could germinate successfully. The number of fungal spores landing in a compartment was assumed to be Poisson‐distributed. The spores were assumed to vary in their sensitivity toward a fungicide, and the sensitivity of the spore population was assumed to be normally distributed around a mean sensitivity. Under these assumptions, the ED 50 value was shown to be a positively biased estimate of the mean sensitivity in the fungal population, and the bias increased with inoculum density. Consequently, the estimated ED 50 value is expected to vary between experiments and laboratories if the inoculation density varies over a considerable range. The correlation between inoculum density and estimated ED 50 value is expected to be strongest when the assayed fungal population consists of several genotypes differing in sensitivity. This expectation was tested by estimating the ED 50 values for one barley powdery mildew isolate at different inoculation densities for three different fungicides. These ED 50 values were positively correlated strongly with the inoculum density for fenpropimorph and more weakly for triadimenol, whereas no correlation was observed for propiconazole.